Jon Gerow · 06-26-2006 · Category:
Testing your power supply's voltage: Software vs. Multimeter
Testing your power supply
with software will typically give incorrect readings. So although the monitoring
software may look fancy and appear to be intricate in its function, it really
can't be considered a valuable diagnostic tool.
There's one chip on the
motherboard that manages health monitoring (typically referred to as the "Winbond
chip" because the chip is typically made by Winbond.) Software gets
it's readings from the same chip the BIOS does, so to say readings from the
BIOS would be more accurate would be a falsehood. The margin of error on
this chip is pretty great because it needs to be taken into consideration
where this chip gets it's data (the ATX connector) and the resistance created
by the traces between the ATX connector and the chip.
Software isn't COMPLETELY
useless: If readings are WAY OFF, there may really be a problem with the
power supply. But always double check the ATX connector to the motherboard
first. If for some reason the connector is not plugged all of the way in
the board, the resistance from the loose connection can cause low voltages.
And never pay any mind
to the -5V or -12V reading in software. Those are typically EXTREMELY off
simply because most power supplies have no -5V lead and most motherboards
don't use the -12V lead. The software simply does not know how to report
back the nonexistent readings and some pretty wacky numbers are often the
To test your power supply
voltages in the most basic of ways you are going to need an electronic device
called a multimeter, (more than likely a Digital Multimeter, or "DMM" as
it's referred to here) which you will physically have to plug into the connectors
on your power supply. Of course, you will need to know what leads are what
on your power supply. Yellow is typically 12V, red is typically 5V and orange
is typically 3.3V. If you put your multimeter's red probe to the tip of any
one of these leads, and the black probe to a black wire (any will do as a power
supply's grounds are all common) then you can get an accurate reading of your
Now keep in mind that not "any multimeter will do" or that testing just "any
drive connector should suffice."
I'm not saying you need a $300 Fluke. There's some $20 DMM's that do the task
just fine. But there's also some $20 DMM's that won't.
Take a look at the "accuracy" of any DMM you're considering using for computer
diagnostics. Even my Fluke is only accurate to 1%, and it only shows digits
to the 1/10 of a volt (.1V resolution.) On the other hand, I had a clamping
multi-meter from a company called "Center" that only sold for $100 that was
accurate to .8% and has a .001V resolution. And to think; I sent my Center
223 to Tulatin and kept the Fluke 336 for myself!
Also, keep in mind one
of the things we learned during the "Power Supply Label" part of this primer.
Different connectors can be on different rails. So if you have a dual-rail
power supply and you probe a peripheral power connector, you're reading a
different rail then what the CPU is getting it's juice from.
Essentially, don't let
software readings scare you. They might not be right. Don't fix what's not
broken. If you are experiencing problems and suspect the power supply is
the source of the problems, check the software, confirm with a decent multimeter
and be thorough!
- The PC power supply:
- The PC power supply label:
- Defining the connectors of an ATX power supply
- ATX power supplies DO NOT turn on at the flip of a switch
- Testing your power supply's voltage: Software vs. Multimeter
- Power Supply Efficiency
- The Derating Curve
- Power Factor Correction
- The resistance of modular connectors, adapters and splitters